JHCXXX10.1369/0022155415589119BaskinHistochemical Identification of Islet Cell Types research-article5891192015

Perspective

Journal of Histochemistry & Cytochemistry 2015, Vol. 63(8) 543­–558 © The Author(s) 2015 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1369/0022155415589119 jhc.sagepub.com

A Historical Perspective on the Identification of Cell Types in of Langerhans by Staining and Histochemical Techniques

Denis G. Baskin Veterans Affairs Puget Sound Health Care System, Research and Development Service, Seattle, WA, USA (DGB); Department of Medicine, Division of Metabolism, and Nutrition, University of Washington, Seattle WA, USA (DGB)

Summary Before the middle of the previous century, cell types of the pancreatic islets of Langerhans were identified primarily on the basis of their color reactions with histological dyes. At that time, the chemical basis for the staining properties of islet cells in relation to the identity, chemistry and structure of their hormones was not fully understood. Nevertheless, the definitive islet cell types that secrete glucagon, insulin, and somatostatin (A, B, and D cells, respectively) could reliably be differentiated from each other with staining protocols that involved variations of one or more tinctorial techniques, such as the Mallory- Heidenhain azan trichrome, chromium hematoxylin and phloxine, aldehyde fuchsin, and silver impregnation methods, which were popularly used until supplanted by immunohistochemical techniques. Before antibody-based staining methods, the most bona fide histochemical techniques for the identification of islet B cells were based on the detection of sulfhydryl and disulfide groups of insulin. The application of the classical islet tinctorial staining methods for pathophysiological studies and physiological experiments was fundamental to our understanding of islet architecture and the physiological roles of A and B cells in glucose regulation and diabetes. (J Histochem Cytochem 63:543–558, 2015)

Keywords diabetes, beta cells, glucagon, immunocytochemistry, immunohistochemistry, insulin, islet cells, , staining, somatostatin

The islets of Langerhans were discovered in 1869 by Paul that they may produce internal secretions that regulate gly- Langerhans when he was a medical student at the Friedrich cemia (Laguesse 1893). Wilhelm University in Berlin (Fig. 1). A student of the emi- The present article is a retrospective history of the histo- nent pathologist Rudolf Virchow, Langerhans described the logical and histochemical staining methods that have been microscopic anatomy of the rabbit pancreas in his M.D. the- used by anatomists and pathologists over the years to iden- sis and reported the presence of “…small cells of almost tify hormone-secreting cell types of the islets of Langerhans perfect homogeneous content and of a polygonal form, with (hereinafter called islets) and understand their function in round nuclei, mostly lying together in pairs or small groups” glucose homeostasis and the pathophysiology of diabetes (English translation) (Sakula 1988). The function of these cells was, of course, unknown to Langerhans (although he Received for publication December 18, 2014; accepted April 15, 2015. suspected that they might be neural in nature) and, except for describing their morphology, he did not give them a Corresponding Author: name. The term “islets of Langerhans” was introduced in Denis G. Baskin, VA Puget Sound Health Care System, Research & Development Service, 1660 S. Columbian Way, Mail Stop S-151, Seattle, 1893 by Edouard Laguesse, who observed them in the WA 98108, USA. human pancreas and (with remarkable foresight) suggested E-mail: [email protected] 544 Baskin

relationship of islet cells to a rich capillary network sug- gested that these cells secrete a substance into the blood to influence carbohydrate metabolism (Laguesse 1893; Diamare 1899; Schäfer 1895), a hypothesis that required evidence of physiological independence of the islets from the exocrine cells of pancreas. Debate centered on the question of whether the islets represented degranulated pancreatic exocrine cells, as it had been observed that pancreatic exocrine cells that were “exhausted” by alkaloid treatment resembled islet cells. Researchers soon discovered that removing the pancreas pro- duced elevated blood sugar and diabetes in experimental ani- mals (von Mering and Minkowski 1890; Minkowski 1893). Pathologists, notably Eugene Opie, found lesions of the islets in that were removed at autopsy from people who were afflicted with diabetes (Opie 1901a, 1901b), thus mak- ing a link between diabetes and a deficiency in an islet secre- tion (that was later named insulin) that was eventually isolated and used to treat patients with type 1 diabetes two decades later (Bliss 2007). In the years that followed, we have learned a great deal about the diversity and function of the cells that comprise these miraculous little endocrine organs, which are vital to life, have so many tasks that are essential for metabolic Figure 1. Paul Langerhans 1878. homeostasis, and whose failure can result in diabetes. How we came to understand the identities and functions of the cell types that comprise the islets is a fascinating story of mellitus (hereinafter called diabetes). The main theme of how microscopic anatomy and staining methods comple- this account focuses on the cells that secrete the canonical mented the functional analysis of islet cell physiology and islet hormones—insulin, glucagon, somatostatin, and pan- the pathophysiology of diabetes in humans and experimen- creatic polypeptide—recognizing that other endocrine fac- tal animals. Much of what we know about their function has tors may also be expressed in the islet, and that neural been learned from microscopic studies of the pancreas (Ahrén et al. 2007), extracellular matrix (Westermark and using chemicals that “stain” islet endocrine cell types, per- Westermark 2013), and stromal (Bollyky et al. 2012) ele- mitting them to be visually distinguished from each other ments are also essential components of the functioning islet. and from pancreatic exocrine cells, neural tissue, and stro- Primarily for convenience, I make an arbitrary distinc- mal elements. tion between the terms “tinctorial staining” (i.e., histologi- cal staining methods that basically reveal microscopic Nomenclature of Islet Cells anatomy) and “histochemical staining” (methods that iden- tify chemical constituents of cells and organs). Tinctorial Over the years, the naming of islet cell types has followed and histochemical staining methods both impart contrast three different but parallel conventions. Many authors use (most often as colors) to islet cells, including their intracel- the Greek letter designations “α cell” and “β cell” when lular secretory granules, and are useful for interpreting the referring to the cells that produce glucagon and insulin, microscopic anatomy of islets. Admittedly, tinctorial vs. respectively, whereas others spell out “” and “beta histochemical is a somewhat arbitrary distinction, as even cell”, and many papers in the literature use the Roman letter the classical tinctorial methods for staining different islet terms “A cell” and “B cell” for these same cell types. cell types are grounded in differences in the chemical prop- Following the latter convention, the islet cell that secretes erties of the respective hormones (or other components of somatostatin is called the “D cell” (also δ or ) and their cytoplasmic granules); although, these properties were the islet cell that secretes is called (and, in some cases, still may be) unknown. the “F cell” (also called the PP cell). The trend in recent years seems to favor the Greek letter names for islet cells, Islet Cells and Diabetes especially for the cells that secrete glucagon and insulin, although it is understood that the Greek and Roman terms By the end of the 19th century, experimental pathologists and are interchangeable in this context. For uniformity and clar- physiologists had hypothesized that the intimate anatomical ity, the present article uses the Roman letter terms for islet Histochemical Identification of Islet Cell Types 545 cell types, regardless of the convention that may have been Lane’s key observation was that fixation in alcohol- originally used by authors in the cited articles. chrome-sublimate resulted in the staining of a morpho- logically different population of islet cells than was stained Early Staining Methods for Islet Cells following fixation in aqueous-chrome-sublimate. Lane inferred that the guinea pig islet contained two chemically Pioneering anatomical studies of islets were performed distinct types of cells, a type with granules that contain primarily with histological stains such as hematoxylin and a chemical that is fixed (precipitated) with alcohol and a eosin, which permitted observations of changes in islet type that contains granules containing a chemical that is morphology but had minimal value for the identification fixed with the aqueous-chrome-sublimate solution. The or discrimination of islet cell types. The problem was former Lane called A cells and the latter he named beta addressed in a classic paper by Lane in 1908: “The princi- cells (Fig. 2) (later changed to “B cells” by Bensley 1911, pal difficulty thus far in dealing with the Islets of 1914), which we now know secrete glucagon and insulin, Langerhans has been the want of a definite method by respectively. Prophetically, Lane suggested that the islets which to distinguish the cells of the islets from the cells of “….in all probability have the function of producing a the pancreas itself; for although there is an apparently con- twofold substance which, poured into the blood stream, stant content of islet tissue in the pancreas, and although has an important effect on metabolism.” Lane’s seminal the areas of islet tissue, in sectioned pancreas, stand out in discovery was that the chemical nature of the granules of sharp contrast with the tubules of the pancreas [Note: this these cell types differed and that this difference could be means the exocrine acini], the physiological distinctness distinguished by fixation and staining, arguing that the of one kind of tissue from the other is the very question respective cells had separate functions and that both were upon which histologists and pathologists have most dis- chemically distinct from each other and from exocrine cell agreed” (Lane 1908). zymogen granules. Lane’s finding that A and B cells can By the beginning of the 20th century, microscopists had be histochemically distinguished by alcohol fixation is recognized two types of islets cells based primarily on mor- consistent with modern immunohistochemical staining phology but also on staining properties with safranin and observations that islet B cells immunostain with insulin gentian violet (Diamare 1899; Laguesse 1895; Schultze antibodies very poorly in pancreas fixed in ethanol, 1900; DeWitt 1906). These dyes are the components of the whereas A cells immunostain robustly with glucagon anti- classic Gram stain, which identifies Gram-positive bacteria bodies after fixation in ethanol (Baskin 2014), and fits based on the chemical and physical properties of their cell with the well-known fact that ethanol extracts insulin from walls; although, the functional meaning of their staining of the pancreas. nuclei and cytoplasmic granules of islet cells was unknown. The papers on islet staining from this early era provide However, it seems likely that these two types of cells stained fascinating reading into the scientific mindset of the authors. with safranin and gentian violet corresponded to A and B The preparation of tissue and the staining solutions as well cells (Bloom 1931). as the protocols are meticulously detailed. The hand-drawn illustrations are works of intricate detail and art, revealing Beginnings of Islet Histochemistry that these investigators were superb microscopists and observers (see Figs. 2, 3). The descriptive narratives are The first bona fide histochemical differentiation of islet often written in the first person, giving the accounts a sense cell types can probably be attributed to Lane (Lane 1908), of immediacy and the reader a feeling of listening to the who identified two types of granular cells in islets based on author. Sadly, such diction and tone are rare in today’s pub- their different solubilities in alcohol. Lane derived this lishing environment. conclusion from the staining characteristics of paraffin- Despite the elegant work of Lane, for the next 30 years, embedded sections of guinea pig pancreas after treatment there was little consensus on the identification of islet cell with one of three different fixative solutions: (a) alcohol- types and their functions. Likewise, it was still unknown chrome-sublimate (an alcoholic solution of potassium whether the postulated cell types that had been identified in dichromate and mercuric chloride), (b) aqueous chrome- guinea pig and human pancreas were representative of sublimate (an aqueous solution of potassium dichromate other species. Whereas the work of Lane showed that it and mercuric chloride), and (c) 70% alcohol. [Note: was possible to differentiate A and B cells using chrome although sometimes not specified in early papers, “alco- sublimate staining and solubility in alcohol, the technique hol” usually meant ethanol.] The sections were stained also left some islet cells unstained, especially in guinea pig with Bensley’s neutral gentian violet, an alcoholic solution islets. Bensley called these unstained cells “clear” or “C that stains the cytoplasmic granules of exocrine and islet cells” and concluded that they represented undifferentiated cells a deep violet, and Orange G, an azo dye that stains the islet cells (Bensley 1914). Later, Bloom recognized three cytoplasm orange (Bensley 1914). types of granular cells in human islets that had been fixed 546 Baskin

Figure 2. Hand-drawn figures from “The cytological characters of the areas of Langerhans,” by Lane, American Journal of Anatomy, Volume 7, Issue 3, Pages 409–422, 10 November 1907, illustrating guinea pig islets stained with Bensley’s neutral gentian following different fixations. Panel A represents an islet from a pancreas fixed in 70% alcohol. The islet A cells have a violet-stained cytoplasm, whereas the islet B cells are essentially unstained. Panel B represents an islet from a pancreas fixed in aqueous chrome-sublimate (no alcohol) and shows islet B cells filled with minute granules that stained violet, whereas the islet A cells are stained light orange. In panel B, some pancreatic acinar cells with clumps of violet-stained granules are depicted at the edge of the islet. Copyright John Wiley and Sons. Published with permission.

in Zenker-formol (mercuric chloride, potassium dichro- mate, sodium sulfate, and formalin) (also called Helly’s fixative). When the islets were stained with the Mallory- Heidenhain azan trichrome technique, Bloom observed A and B cells as described by Lane and also a third granu- lated cell type in human islets that he called the D cell, but he was not able to identify a clear (C) cell in human islets (Bloom 1931) (Fig. 3). In general, these tinctorial staining methods produced variable results on the islets of different species and it was difficult to correlate specific cell types with specific func- tions. Interestingly, in 1937, Thomas reported a study of islet staining from the tail of the pancreas of 41 different mammalian species representing ten orders and 25 fami- lies (including bats, ring-tailed coati, flying squirrel, camel, and jaguar). He claimed that Lane’s fixation and staining method produced inferior and inconsistent results as compared with other techniques. Thomas used more than seven different staining techniques in his study, although the most useful was said to be the Mallory- Figure 3. A drawing by Bloom from “A new type of granular cell in Heidenhain azan trichrome method. He reported that the the islets of Langerhans of man,” from The Anatomical Record: Advances staining reaction of all the mammalian islet species stud- in Integrative Anatomy and Evolutionary Biology, 1931, illustrating a human ied was essentially similar to the description that had islet stained with the Mallory-Heidenhain azan trichrome technique. A cells (A) are stained reddish brown, D cells (D) are blue, and B already been established for the human pancreas (Bloom cells (B) are stained faintly orange. The figure shows pancreatic acinar 1931). He also identified three types of granular cells cells (Pac) and reticular fibers (Ret). Copyright John Wiley and Sons. (called A, B, and D cells) on the basis of granularity and Published with permission. differential coloration of cytoplasmic granules in all 41 Histochemical Identification of Islet Cell Types 547

technique in his 1941 paper, based on 70 normal and patho- logical human pancreases (Gomori 1941). Following the protocol of the Mallory-Heidenhain azan trichrome method, the sections were oxidized with potassium permanganate, decolorized with sodium bisulfite, and then stained in an acidic potassium dichromate hematoxylin solution. The reaction was monitored with a microscope until the desired staining was obtained; then the sections were differentiated in acid alcohol. The counterstain was phloxine, a common dye that renders the cytoplasm a colorful reddish hue. When properly applied, this technique—known as the chromium hematoxylin-phloxine method—stained the islet B cells an intensely deep blue and the A cells bright red. The D cells stained capriciously and variably pink-to-red in hue and were difficult to distinguish from islet A cells on the basis of color. In contrast, with the Mallory-Heidenhain azan tri- chrome stain, D cell granules could be stained deep blue, contrasting with red A cell granules; but the B cell granules (and thus the B cells) appeared unstained. Thus, by using both the Mallory-Heidenhain azan trichrome and the chro- mium hematoxylin-phloxine stains on adjacent sections, an investigator could get crude (but reproducible) differential counts of A, B, and D cells from pancreas sections and, moreover, observe how the numbers and appearances of these cell types were associated with pathophysiological Figure 4. George Gomori. Image courtesy Special Collections changes in islet function in humans and experimentally in Research Center, University of Chicago Library. Published with animal models. permission. The chromium hematoxylin-phloxine method (which became known as the first “Gomori stain”), paired with the species, although Bensley’s unstained C cell was found Mallory-Heidenhain azan trichrome stain, was a standard only in islets of guinea pig and opossum. approach for identifying islet A, B, and D cells for several decades. It was almost always used following formalin fixa- The Gomori Age tion, generally Bouin’s fixative or Zenker-formol. Although it stains islet B cells with an intensity that is in some mea- Before the 1930s, the stains used to study islet A and B cells sure proportional to insulin content, the chromium hema- gave excellent cellular differentiation on the islets of guinea toxylin-phloxine method is not biochemically specific for pig pancreas but relatively poor and irregular results on the insulin. Moreover, the chromium hematoxylin-phloxine islets of other species including humans. The Mallory- technique also stains many extrapancreatic cell types and Heidenhain azan trichrome method, a useful stain for islet A was widely used, for example, to stain zymogen granules cells, unfortunately stained islet B cells rather poorly red in pancreatic exocrine cells and adenohypophyseal cell (Bloom 1931). There was a strong need for robust staining types (prolactin and growth hormone cells stained red, thy- methods to identify islet B cells in particular to complement rotropes and gonadotropes stained blue). When applied to the growing field of islet pathophysiology in animal models islets, the resulting blue stain was considered to be specific as well as for human diabetes. Watershed events of this era for B cells. were the publication of two methods for staining islet B cells by George Gomori, a professor of medicine at the Aldehyde Fuchsin Method University of Chicago (Fig. 4). In 1950, Gomori published an aldehyde fuchsin staining Chromium Hematoxylin-phloxine Method technique for elastic fibers that also stains cytoplasmic granules of islet B cells as well as a variety of other endo- In 1939, Gomori published a staining protocol that sharply crine cell types, including adenohypophyseal basophils, distinguished B cells from other islet cell types in human neurohypophyseal neurosecretory cells, mast cells, and gas- pancreases that were fixed in Bouin’s fluid (picric acid, for- tric chief cells (Gomori 1950). This protocol, known as the malin, acetic acid) (Gomori 1939). Gomori refined this “Gomori aldehyde fuchsin stain”, is capable of producing 548 Baskin

Figure 5. Rat islet stained with Gomori’s aldehyde fuchsin after oxidation in periodic acid. Image from “A comparison of the staining affinities of aldehyde fuchsin and the Schiff reagent,” by Scott and Clayton, Journal of Histochemistry and Cytochemistry, 1953. Copyright Histochemical Society. Published with permission. Figure 6. Rat islets stained with Gomori’s aldehyde fuchsin intense staining of islet B cells (Fig. 5). However, the tech- method. Top panel is an islet from a control rat, showing nique produced variable staining results among different intensely stained granulated B cells. Bottom panel shows an islet from rat that was treated with a sulfonylurea to stimulate insulin investigators and species. The problem of inconsistent secretion, showing depletion of staining. Image from “A portrait results was attributed to the mode of preparation, storage, of the pancreatic B cell,” by Orci, Diabetologia, 1974. Published and stability of the aldehyde fuchsin solution, which was with kind permission kind permission from Springer Science and prepared from basic fuchsin and allowed to “age” before Business Media. use, and also to the variety of fixative solutions that were used by different investigators (Cameron and Steele 1959; disulfide bonds to sulfonic acid groups, which act as decol- Mowry et al. 1980; Mowry and Kent 1988). Despite these orized Schiff reagents. In the presence of aldehyde, the difficulties, the method had appeal partly because the inten- uncolored Schiff reagent changes to a magenta color, thus sity of islet B cell staining with Gomori’s aldehyde fuchsin presumably staining insulin in B cell granules (Bangle appeared to be roughly related to islet insulin content (Fig. 6) 1954, 1956; Bangle and Alford 1954). and other islet cell types were unstained. Gomori’s alde- The feasibility of this mechanism was tested by biochemi- hyde fuchsin method was widely used in various formula- cal studies on the reaction of aldehyde fuchsin with insulin by tions (Scott 1952; Scott and Clayton 1953; Cameron and Kvistberg et al. (1966), who analyzed the staining of beef zinc Steele 1959; Mowry 1983) until immunohistochemical insulin in polyacrylamide gels following disc electrophoresis. methods supplanted routine use of tinctorial B cell stains. They prepared aldehyde fuchsin according to Gomori’s recipe and aged it for 3 days before use (the “age” of aldehyde fuchsin Chemistry of Aldehyde Fuchsin Staining of Islet solutions was found to affect its staining properties, although B Cells the reasons for this were not understood). The gels were oxi- dized with KMnO and H SO before staining (control gels 4 2 4 The chemical basis for the method was assumed to be the were unoxidized), and then stained in the aldehyde fuchsin reaction of aldehyde fuchsin with insulin after prior oxida- solution, and subjected to destaining to remove unreacted dye. tion by KMnO or periodic acid (Scott 1952). Following the This was essentially the same protocol that was used on pan- 4 elucidation of the chemical structure of insulin, Scott and creas tissue sections. The authors observed that aldehyde fuch- Clayton (1953) hypothesized that insulin is oxidized at sin stained insulin in the gels only if they had been oxidized Histochemical Identification of Islet Cell Types 549

(Barrnett and Seligman 1952b; Barrnett 1953) and disulfide groups (Barrnett and Seligman 1952a, 1954) in tissue sec- tions. Barrnett and Seligman, recognizing that insulin is rich in disulfide owing to its 12% cysteine content (Sanger and Tuppy 1951a, 1951b), used purified crystalline insulin in experiments to develop a histochemical method for staining insulin and used physiological experiments to demonstrate its validity (Barrnett et al. 1955). Barrnett and Seligman fixed pancreases from rabbits, albino rats, mice, dogs, toad- fish, and humans in formalin-based solutions, including Bouin’s fixative, and embedded tissues in paraffin. Gomori’s aldehyde fuchsin method was used to confirm that the histo- chemical staining for sulfhydryl and disulfide groups indeed identified islet B cells. In carefully controlled experiments that included measurements of plasma glycemia and altered pancreatic insulin content after cytotoxic destruction of islet B cells with alloxan, they observed parallel changes in islet B cell staining for sulfhydryl/disulfide groups and islet insu- lin secretion. Moreover, islet B cell staining was abolished when the pancreases were fixed in acid ethanol, which extracted insulin from the pancreas (Barrnett and Seligman Figure 7. Reproduction of Figure 1 from “Staining of insulin 1954) (Fig. 8). Although not uniquely specific for insulin per with aldehyde fuchsin,” by Kvistberg, Lester, and Lazarow, se (as cysteine sulfhydryl groups are present in many pro- Journal of Histochemistry and Cytochemistry, 1966, demonstrating staining of gels after disk electrophoresis. (A) Beef insulin stained teins), the Barrnett and Seligman histochemical method was with a dye for protein (aniline blue black) immediately after gel considered specific for B cells within islets (supported by electrophoresis. (B) Beef insulin stained with Gomori’s aldehyde the later finding that glucagon lacks disulfide bonds). Its fuchsin after gel was oxidized with KMnO -H SO . The aldehyde 4 2 4 widespread adoption for islet studies was nevertheless hin- fuchsin stained the same band that contains insulin in gel A. (C) dered by its technical difficulty compared with the more Beef insulin that was stained with aldehyde fuchsin without prior conventional Gomori aldehyde fuchsin method, which KMnO -H SO oxidation, showing a lack of stained insulin band 4 2 4 remained the method of choice for most investigators until in the absence of oxidation. Copyright Histochemical Society. Published with permission. tinctorial methods were eclipsed by immunohistochemical methods for identifying islet cell types. after electrophoresis (Fig. 7). The authors confirmed these results by eluting insulin from unstained gels with acid alcohol Pseudoisocyanin Methods and assaying for insulin by immunoassay. This study, by Kvistberg et al. (1966), is a classic demonstration of the appli- An alternate histochemical technique for staining islet B cation of biochemical techniques to understand the chemical cells, based on the metachromatic reaction of oxidized insu- basis of histochemical staining specificity of a dye molecule lin with pseudoisocyanin, was described by Coalson (1966). for insulin in islet B cells. Later, Greenwell et al. (1983) per- In this protocol, the three cysteine disulfide bonds of insulin formed a controlled analysis on the effects of fixation and oxi- are oxidized to sulfonic acid groups, which are subsequently dation on the ability of aldehyde fuchsin to stain insulin, detected with pseudoisocyanin dyes, using a method pub- proinsulin, and other proteins in polyacrylamide gels. These lished by Schiebler and Schiessler (1959). The reaction latter investigators confirmed that the oxidation treatment was requires the sections to be oxidized in strong sulfuric acid necessary to obtain positive aldehyde fuchsin staining but also and potassium permanganate, which also considerably concluded that this staining reaction was not related to the cys- damages tissue structure. Coalson demonstrated that the teine content of the proteins, thus raising unresolved doubt resulting deep purple metachromatic staining was localized about whether either insulin or proinsulin is actually responsi- to islet B cells. Despite its apparent specificity for B cells in ble for aldehyde fuchsin staining of islet B cells. the islets, the pseudoisocyanin staining method failed to attract much use. This may have been because, although the staining it produced was comparable to that obtained with Barrnett and Seligman Technique Gomori’s aldehyde fuchsin method (Fig. 9), the Gomori The efforts to base islet B cell staining on the chemical struc- aldehyde fuchsin protocol was already in common use and, ture of insulin benefitted from developments in colorimetric more significantly, the metachromatic staining produced by methods for detecting protein-bound sulfhydryl groups pseudoisocyanin was unstable and usually faded overnight. 550 Baskin

Figure 8. Staining of islet B cells in paraffin sections of rat pancreas for sulfhydryl and disulfide groups. Left panel shows absence of staining in a rat islet after extraction of insulin by fixation in acid ethanol. Right panel shows positively stained B cells after fixation in Zenker’s fluid (mercuric chloride, potassium dichromate, sodium sulfate, acetic acid). Image from “Histochemical demonstration of sulfhydryl and disulfide groups of protein,” by Barnnett and Seligman, Journal of the National Cancer Institute, 1954.

Islet “Third Cells” 1967; McGadey 1979; Schweisthal et al. 1981) as well as by silver impregnation techniques (Hellerstrom and By the end of the 1950s, and before the adoption of immu- Hellman 1960; Hellman and Hellerstrom 1961; Grimelius nohistochemical staining techniques, islet A and B cells had 1968a; Fujita 1968). Bowie’s gamma cell was identified in been established as functionally distinct cell types secreting fish islets and is probably not equivalent to the mammalian glucagon and insulin, respectively. It was also established D cell, which has also been called a gamma cell (Mosca that islet A and B cells could be identified microscopically 1957); although, the latter term has also been used for the by specific tinctorial staining methods, and true histochemi- islet pancreatic polypeptide (F or PP) cell (Malaisse-Lagae cal stains, especially for islet B cells, were incipient. et al. 1977). On the basis of staining reactions, the “third However, cells that were distinguishable from A and B cells cell type” in islets of mammals as well as fish was inter- by virtue of different staining and morphological character- preted by many early investigators as a functional stage of istics had also been described in islets of various vertebrate either A or B cells. This literature can be confusing in retro- species (Bensley 1911; Bowie 1925; Bloom 1931; Mosca spect because of the uncertainty about the homology of 1957), leading to the question of whether islets contained a “third cell types”, given the different names used by differ- functional “third cell type”. The term “third cell type” was ent investigators and in different species, and because the applied to any epithelioid islet cell that did not show the term is no longer in use. staining characteristics of A or B cells with classic tinctorial techniques. A variety of terms were used to name these Argyrophilic Islet Cells non-A and non-B islet cells, such as gamma cells (Bowie 1925), C cells (Bensley 1911), D cells (Bloom 1931), and A silver impregnation staining technique for identifying and delta (or δ cells (Schweisthal et al. 1981). In general, islet classifying islet cell types was developed by Scandinavian “third cell” types were chromophobic (i.e., had clear cyto- histochemists in the 1960s (Hellerstrom and Hellman 1960, plasm or stained weakly with conventional tinctorial meth- Hellman and Hellerstrom 1961; Grimelius 1968a, 1968b). ods), presumably because of their sparse granule content, as The silver staining methods involved deposition of metallic seen with conventional brightfield light microscopy. silver over cells that were able to reduce silver nitrate to However, these cells could, in some cases, be stained by metallic silver after exposure to a reducing agent such as certain tinctorial methods such as the Mallory-Heidenhain hydroquinone or formaldehyde. This property is called argyr- azan trichrome stain (e.g., Thomas 1937) and others (Epple ophilia, and cells that stain with these methods are called Histochemical Identification of Islet Cell Types 551

Figure 9. Staining of B cells in rat islet with pseudoisocyanin. Left panel shows a paraffin-embedded section stained with pseudoisocyanin, with intense staining of B cells. Right panel shows the same islet as that in the left panel after the pseudoisocyanin stain was removed from the section, and depicts B cells after the section was restained with Gomori’s aldehyde fuchsin. Image from “Pseudoisocyanin Staining of Insulin and Specificity of Empirical Islet Cell Stains,” by Coulson, Stain Technology, 1966. Copyright Informa Healthcare. Published with permission.

properties with A cells. Grimelius called them α cells and α 1 2 cells, respectively, using terminology that Hellman et al. (1962) had proposed for dog islets. The selectivity of the technique was dependent on formaldehyde fixation, as fixa- tion in ethanol-H S caused all islet cells to stain black 2 (Grimelius 1968b). Wilander and Westermark (1976) later used electron microscopy to demonstrate that the distinctive cytoplasmic granules of A and D cells exhibited correspond- ingly different patterns of silver deposition, confirming that the argyrophilic α cells were equivalent to A cells and the 2 argyrophilic α cells were D cells. The D cell was eventually 1 validated as a unique islet cell type that secretes somatostatin by electron microscopy and immunohistochemical staining (Luft et al. 1974; Orci et al. 1975; Erlandsen et al. 1976). The silver staining properties of α cells and α cells were 1 2 Figure 10. Islet from a human pancreas stained with the silver interpreted to mean that these cells were also capable of nitrate technique, showing black (argyrophilic) islet cells against secreting biogenic amines (Falck and Hellman 1963; Cegrell unstained background. Image from “The argyrophil reaction in 1968). This was based in part on the similarity of α cell and islet cells of adult human pancreas: studies with a new silver nitrate 1 α cell argyrophilic staining properties to those of intestinal procedure,” by Grimelius, Acta Societatis Medicorum Upsaliensis, 1968. 2 Copyright Informa Healthcare. Published with permission. argyrophilic cells that were thought to secrete biogenic amines. This hypothesis was extended by Pearse (Pearse 1969; Pearse and Polak 1971) to include the pancreatic islets argyrophilic cells. In the islets, argyrophilic cells appear in his APUD (amine precurser uptake and decarboxylation) intensely black against a yellow background that includes cell concept. According to this hypothesis, APUD cells unstained B cells (Fig. 10). Grimelius (1968a, 1968b) identi- shared the capability of synthesizing biogenic amines and fied two cell types that could be distinguished on the basis of were proposed to have a common developmental origin from their argyrophilic staining properties in formalin-fixed paraf- neural crest ectoderm. The APUD concept included entero- fin sections of human autopsy specimens: One type had an endocrine argyrophylic cells that secreted peptides as well as elongated shape and stained intensely black with the argyro- many other neuroendocrine cell types that are not argyro- philic reaction; another more numerous cell type showed philic. Thus, both islet A cells (i.e., the argyrophilic α cells 2 lighter argyrophilic staining and shared tinctorial staining described by Grimelius) and islet B cells (that are 552 Baskin

not argyrophilic) were considered to have the potential of vital stain that has proved useful for islet sorting is a fluores- secreting biogenic amines by Pearse. The islet D cell (then cent dye known as TSQ (Jindal et al. 1993), a toluene sulfon- called the α cell) was also considered to be an APUD cell by amide that also binds to zinc. More recently, cultured islets 1 Pearse based on its argyrophilic properties as well as reports have been labeled in vitro with dextran-coated superpara- that it contained the enteroendocrine hormone gastrin (this magnetic iron oxide (SPIO) nanoparticles and subsequently was before somatostatin had been discovered). It’s now visualized in vivo by magnetic resonance imaging after trans- accepted that the APUD concept is invalid at least to the plantation into the liver and kidney capsule (Evgenov et al. extent that it included islet B cells, which appear not be 2006a, 2006b; Zacharovová et al. 2012). The iron oxide con- derived from neuroectoderm (Andrew 1976; Pictet et al. trasting agents appear to label all islet cell types and the stain- 1976). Furthermore, the specificity of the argyrophilic stain- ing can be visualized in tissue sections by light microscopy ing technique for identifying cells that secrete biogenic (Evgenov et al. 2006a). amines does not agree well with subsequent pharmacological Cobalt was shown to be concentrated in pancreatic islet and physiological studies. For example, although islet B cells tissue of a marine teleost fish by Sture Flakmer (Falkmer do not stain with the classic argyrophilic staining methods, et al. 1964) using a histochemical technique that involved in recent evidence suggests that islet B cells secrete GABA, situ precipitation of metals as insoluble sulfides. Similar to dopamine, and serotonin (Garry et al. 1986, 1988; Ohta et al. zinc, cobalt has the property of crystallizing with insulin and 2011; Ustione and Piston 2012). Despite these shortcomings it forms complexes with sulfhydryl groups such as cysteine of the interpretations of islet cell argyrophilic staining, it’s and glutathione. These experiments were facilitated by the fair to say in retrospect that the method was very useful for fact that teleost islets are gathered into discrete lobes called identifying islet A and D cells in the pancreas of many verte- Brockmann bodies. Several days after injection of cobalt brate species before the advent of immunohistochemistry. chloride into the living fish, the islet lobes were processed for histochemical detection of cobalt. This involved expos- Zinc and Cobalt Techniques ing the islets to hydrogen sulfide gas prior to fixation and embedding. The resulting metal sulfides were then demon- Islet B cells contain a high level of zinc, which plays a key strated with a silver-sulfide reaction that produced a black role in the biosynthesis, granular storage, and secretion of precipitate over the islet cells. A comparison with aldehyde insulin. Autoradiographic studies with radioactive zinc fuchsin staining showed that mostly B cells were identified showed that living islet B cells preferentially accumulate zinc with this technique; the islet α cells (equivalent of the argyr- 1 compared with other islet cells (Wagner et al. 1981). This ophilic D cells) were also stained, whereas the islet α cells 2 property proved useful for islet vital staining techniques (equivalent of A cells) did not. Besides not being easily (Okamoto 1942; McNary 1954). Lukowiak developed a adapted to routine physiological and pathological tissue, the zinc-sensitive fluorescent probe based on Newport Green to cobalt technique also had questionable specificity, as others selectively label living islet B cells in human islet cell prepa- also reported that cobalt stained A cells in fish (Mosca 1957). rations for imaging by confocal microscopy (Lukowiak et al. 2001) and fluorescent activated cell sorting (Liew et al. 2008). The zinc-chelating molecule dithizone selectively Emergence of Immunohistochemical stains B cells because of their elevated zinc content. It is used Techniques for Islets to identify B cells in clusters of developing embryonic pan- Immunofluorescence Methods creatic stem cells (Shiroi et al. 2002; Hefei et al. 2015) and also to identify islets during their isolation from whole pan- Shortly after the immunofluorescence histochemical tech- creas for in vitro studies or transplantation (Gray et al. 1983; nique was introduced by Coons and Kaplan (1950), Lacy and Bonner-Weir et al. 2000). Interestingly, Bensley (1911) had Davies (1957) published a method for staining B cells in rat demonstrated that pancreatic islets could be stained with pancreatic islets with fluorescein-labeled anti-insulin immune aqueous neutral red solutions injected via the aorta after serum that was produced in guinea pigs (Fig. 11). They used exsanguination. This technique was adapted to stain living small pieces of mouse and beef pancreas that were frozen in islets for transplantation without impairing their viability or liquid nitrogen and freeze-dried, followed by embedding and insulin secretion by injecting 2% solutions of neutral red into sectioning in wax. Papers describing the immunofluorescent the aorta or i.v. (Gray et al. 1983). With this method, the islets localization of insulin (Lacy and Davies 1959) and glucagon of dogs, pigs, and rats can be seen as darkly staining bodies (Baum et al. 1962) in the islets of humans and other mam- against a pale pink background with a stereomicroscope and malian species soon followed. Nevertheless, the use of are visible in frozen sections of rat pancreas. Although neu- immunofluorescent techniques for staining islet cells was tral red enjoys wide use as a nonspecific, vital stain for many slow to be widely adopted. Few investigators had the exper- cell types, the reason for its specificity for islets following tise to produce specific antibodies or had access to the fluo- intravascular injection is not known. Another islet-specific rescence microscopes needed for this method. Indeed, as late Histochemical Identification of Islet Cell Types 553

investigators no longer had to rely on capricious tinctorial histological stains and relatively nonspecific histochemical methods to identify islet cell types. Nevertheless, novel tinctorial methods for staining islet cells continued to appear occasionally. Kito and Hosoda (1977) described a triple labeling tinctorial protocol that incorporated argyrophilic, aldehyde fuchsin, and lead- hematoxylin stains in a single tissue section. McGadey (1979) developed a method for exquisitely staining A, B, and D cells in the same islet with a combination of alcian blue, chrome hematoxylin, acid fuchsin, and aurantia dyes. Even as late as 1982, the dye orcein was shown to specifi- cally stain B cells in human pancreas (Callea and Desmet 1982), but these esoteric tinctorial methods never gained wide use because of the advantages and simplicity of immu- nohistochemistry for identifying islet cell types. Figure 11. Original publication of immunofluorescent staining of islet B cells for insulin. Image from “Preliminary Studies on the Electron Microscopy Demonstration of Insulin in the Islets by the Fluorescent Antibody Technic,” by Lacy and Davies, Diabetes, 1957. Copyright American Researchers had described the ultrastructural morphology of Diabetes Association. Published with permission. islet cells since the early days of electron microscopy (Lacy 1957; Williamson and Lacy 1959; Meyer and Bencosme as 1966, Coalson stated in reference to the Lacy and Davies 1965). These and later studies established the morphological fluorescent antibody technique for identifying B cells, “…. features of islet cell types, including the structure and fluorescent antibody procedures are too complicated for rou- appearance of their cytoplasmic secretory granules (Orci tine use by many laboratories” (Coalson 1966). 1974). Immunohistochemistry, particularly in combination with ultrastructural morphology provided by transmission electron microscopy, enabled the definitive identification of Immunoperoxidase Methods islet cell types expressing peptides and other molecules that The development of enzyme-labeled antibody techniques could not easily be differentiated with classical tinctorial for immunohistochemistry in the early 1970s revolution- staining methods. Identification of islet cell types by elec- ized islet biology. The methods were relatively simple and tron microscopy was aided by techniques for correlating the reliable, and, perhaps foremost, the results could be visual- analysis of identified individual cells at both the light micro- ized in a conventional brightfield light microscope. The scopic and ultrastructural levels and especially by the appli- commercial availability of labeled secondary antibodies, cation of immunoperoxidase techniques for electron especially soluble peroxidase-antiperoxidase immunoglob- microscopic studies of islets (Pelletier 1977; Erlandsen 1980). Notably, the pancreatic polypeptide secreting F cell ulin G (known as the PAP complex) developed by (also called the PP or gamma cell), generally appears chro- Sternberger and colleagues (1970), brought high quality, mophobic because its secretory granules are not revealed reproducible immunohistochemistry within the reach of all with the classical islet staining methods, whereas the secre- islet investigators. Islet B cells could be stained reliably and tory granules in F cells can be readily stained by immunohis- reproducibly using guinea pig anti-insulin serum (Fig. 12) tochemistry with antibodies to pancreatic polypeptide and that was readily available at that time because these anti- visualized by electron microscopy (Greider et al. 1978; bodies were commonly used for insulin radioimmunoas- Baskin et al. 1984). More recently, a cell that has been called says. Consequently, reports identifying insulin (Misugi the , has been identified at the periphery of et al. 1970), glucagon (Hegre et al. 1976; Erlandsen 1980), human and rat islets using antibodies to ghrelin (Wierup et somatostatin (Orci et al. 1975; Goldsmith et al. 1975; al. 2004; Andralojc et al. 2009); although, its status as a dis- Erlandsen et al. 1976; Hegre et al. 1976), and pancreatic tinct islet cell type remains to be firmly established. polypeptide (Hegre et al. 1976; Greider et al. 1978) based on immunoperoxidase staining, both by light and electron microscopy, soon appeared in the literature, thus confirm- Some Cautions ing the endocrine nature and morphology of the respective In many cases, it is not clear whether the localization of a islet cell types expressing these hormones (Fig. 13). Armed novel molecule to an islet cell represents a functionally with hormone-specific antibodies and both immunofluo- unique cell type (such as the classical A, B, and D cells), as rescent and immunoperoxidase detection methods, islet it is risky to eliminate the possibility of co-expression with 554 Baskin

Figure 12. Immunoperoxidase staining of insulin in islet B cells (B) of guinea pig pancreas. Other islet cells and exocrine (e) cells are unstained. Image from “Immunocytochemical identification of cells containing insulin, glucagon, somatostatin, and pancreatic polypetide in the islets of langerhans of the guinea pig pancreas with light and electron microscopy,” by Baskin, Gorray, and Fujimoto, The Anatomical Record: Advances in Integrative Anatomy and Evolutionary Biology, 1984. Copyright John Wiley and Sons. Published with permission. insulin, glucagon, or somatostatin in the absence of care- account, but I have purposely chosen to omit this technique. fully controlled studies; this has been established, for exam- In situ hybridization, although an extremely powerful histo- ple, by the co-expression of insulin and amylin in islet B chemical technique for identifying gene expression in islet cells (Lukinius et al. 1989) and of chromogranin A with cells and one that can give us a wealth of information about insulin and glucagon in the bovine pancreas (Ehrhart et al. multiple gene expression and the dynamics of gene 1986). In some cases, the expression of markers identified responses to physiological mechanisms, nevertheless has by immunohistochemistry appears to be unique for specific not added much fundamentally to our ability to identify islet islet cell types. For example, Bonner-Weir’s group reported cell types in a physiological context. This thinking could be that the expression of cytokeratin-19, matrix metallopro- accused of being too parochial, as it’s based on a view of teinase-2, and surfactant protein-D were expressed selec- islet cells being characterized by the expression of a single tively in B cells of developing islets and can be considered hormone (e.g., insulin). Admittedly, this is an endocrinolo- as markers of newly formed B cells (Aye et al. 2010). gist’s physiological perspective—islets are classically char- A note of prudence is also appropriate regarding identifi- acterized by cells that secrete the hormones insulin, cation of islet cell types based on immunohistochemical glucagon, somatostatin, and pancreatic polypeptide. An staining for antigens other than the classical islet peptide investigator from another discipline might functionally cat- hormones, especially in the absence of valid controls. Islet A egorize islet cell types based on their expression of recep- cells in particular can show strong nonspecific binding to tors, homeobox genes, transcription factors, immunological IgG. Thus positive immunohistochemical staining for non- markers, or matrix molecules. But I think that an instructive glucagon antigens can be problematic to conclude as local- take home message from this historical review is that, while ized in A cells without rigorous validation controls. This the classic histochemical and tinctorial staining methods phenomenon has been attributed to ionic charge effects, were not necessarily specific for identifying islet hormones hydrophobic interactions, and complement interactions (i.e., Barrnett’s histochemical stains for insulin also stained resulting from basic amino acid residues (and is alleviated sulfide groups in other tissues), nevertheless, within the by use of poly-L-lysine in the incubation buffer) (Scopsi et boundaries of an islet, those techniques could reliably iden- al. 1986; Buffa et al. 1979). The importance of rigorous con- tify A, B, and D cells. trols for standardizing and validating immunohistochemis- Compared to the classical tinctorial and histochemical try has recently been emphasized (Hewitt et al. 2014). methods, antibody-based immunohistochemical staining techniques are unquestionably superior, more reliable and Perspective reproducible, and have a stoichiometry that can be related to hormone content (albeit, crudely). They also allow us to One could argue, and in principle I would not disagree, that co-localize other gene products that are expressed with islet in situ hybridization methods properly belong in this hormones under diverse physiological, pathological, and Histochemical Identification of Islet Cell Types 555

Figure 13. Immunoperoxidase identification of classic islet cell types in rat pancreas. (1) B cells immunostained for insulin. (2) A cells immunostained for glucagon. (3) D cells immunostained for somatostatin. (4) F cells immunostained for pancreatic polypeptide. Image from “Pancreatic islet cell hormones distribution of cell types in the islet and evidence for the presence of somatostatin and gastrin within the D cell, by Erlandsen, Hegre, Parsons, McEvoy, and Elde, Journal of Histochemistry and Cytochemistry, 1976. Copyright Histochemical Society. Published with permission. developmental conditions. While not denying the advan- normal and diabetic subjects were studied and compared tages of antibody-based staining techniques for understand- with staining results). Thus, advances in islet pathophysiol- ing islet cell biology, it is not a stretch to say that A, B, and ogy developed hand in hand with the development of histo- D cells could already be reliably identified on the basis of logical and histochemical staining methods for identifying their morphology and classical tinctorial and staining prop- islet cell types and interrogating their functions. erties before the wide spread use of immunohistochemical techniques. For example, comparing published figures of Acknowledgments islets stained with Gomori’s aldehyde fuchsin method and This material is based upon work supported by the Office of islets immunostained for insulin reveals remarkably similar Research and Development of the Department of Veterans Affairs, images that could be easily seen as identical. and by the Cellular and Molecular Imaging Core of the University In hindsight, it’s clear that the specificity of classical his- of Washington Diabetes Research Center (grant P30 DK017047 tochemical and tinctorial stains for respective islet cell from the NIH National Institute of Diabetes and Digestive and types was not established solely by histological techniques. Kidney Diseases). DGB is the recipient of a VA Senior Research The most seminal of those advances involved staining Career Scientist award. experiments that were conducted in the context of physio- logical and pathophysiological studies (e.g., animals were Competing Interests treated with alloxan to create diabetes, diabetic animals The author declared no potential competing interests with respect were administered insulin, and human autopsy tissue from to the research, authorship, and/or publication of this article. 556 Baskin

Author Contribution insulin, glucagon, somatostatin, and pancreatic polypeptide in the islets of Langerhans of the guinea pig pancreas with light The author is solely responsible for the writing of this article. and electron microscopy. Anat Rec 208:567-578. Baum J, Simons BE, Jr, Unger RH, Madison LL (1962). Funding Localization of glucagon in the alpha cells in the pancreatic The author disclosed receipt of the following financial support for islet by immunofluorescent technics. Diabetes 11:371-374. the research, authorship, and/or publication of this article: This Bensley RR (1911). Studies on the pancreas of the guinea pig. Am material is based upon work supported by the Office of Research J Anat 12:297-387. and Development of the Department of Veterans Affairs, and by Bensley RR (1914). Structure and relationships of the islets of the Cellular and Molecular Imaging Core of the University of Langerhans. Criteria of histological control in experiments on Washington Diabetes Research Center (grant P30 DK017047 the pancreas. Harvey Lectures 1914-1915:250-289. from the NIH National Institute of Diabetes and Digestive and Bliss M (2007). The Discovery of insulin, 25th Anv edition. Kidney Diseases). DGB is the recipient of a VA Senior Research Chicago: University of Chicago Press, p. 310. Career Scientist award. Bloom W (1931). A new type of granular cell inthe islets of Langerhans of man. Anat Rec 49:363-371. Bollyky PL, Bogdani M, Bollyky JB, Hull RL, Wight TN (2012). The References role of hyaluronan and the extracellular matrix in islet inflamma- Ahrén B, Wierup N, Sundler F (2006). Neuropeptides and the tion and immune regulation. Curr Diab Rep 12:471-480. regulation of islet function. Diabetes 55: S98-S107. Bonner-Weir S, Taneja M, Weir GC, Tatarkiewicz K, Song KH, Andralojc KM, Mercalli A, Nowak KW, Albarello L, Calcagno R, Sharma A, O’Neil JJ (2000). In vitro cultivation of human Luzi L, Bonifacio E, Doglioni C, Piemonti L (2009). Ghrelin- islets from expanded ductal tissue. Proc Natl Acad Sci U S producing epsilon cells in the developing and adult human A. 97:7999-800. pancreas. Diabetologia 52:486-493. Bowie DJ (1925). Cytological studies of the islets of Langerhans Andrew A (1976). An experimental investigation into the possible in a teleost Neomcenis griseus. Anat Rec 29:57-73. neural crest origin of pancreatic APUD (islet) cells. J Embrol Buffa R, Solcia E, Fiocca R, Crivelli O, Pera A (1979). Exp Morphol 35:577-593. Complement-mediated binding of immunoglobulins to Aye T, Toschi E, Sharma A, Sgroi D, Bonner-Weir S (2010). some endocrine cells of the pancreas and gut. J Histochem Identification of markers for newly formed beta-cells in the Cytochem 27:1279-80. perinatal period: a time of recognized beta-cell immaturity. J Cegrell L (1968). The occurrence of biogenic monoamines in the Histochem Cytochem 58:369-376. mammalian endocrine pancreas. Acta Physiol Scand Suppl Bangle R Jr (1954). Gomori’s paraldehyde-fuchsin stain. I. 314:1-60. Physico-chemical and staining properties of the dye. J Callea F, Desmet VJ (1982). The demonstration of B cells in pan- Histochem Cytochem 2:62-67. creatic islets with orcein. Histochemical J 14:545-552. Bangle R Jr (1956). Factors Influencing the staining of beta-cell Cameron ML, Steele JE (1959). Simplified aldehyde-fuchsin granules in pancreatic Islets with various basic dyes, Including staining of neurosecretory cells. Stain Technol 34:265-266. paraldehyde-fuchsin. Am J Pathol 32:349-362. Coalson RE (1966). Pseudoisocyanin staining of insulin and spec- Bangle R Jr, Alford WC (1954). The chemical basis of the periodic ificity of empirical islet cell stains. Stain Technol 41:121-129. acid Schiff reaction of collagen fibers with reference to peri- Coons AH, Kaplan MH (1950). Localization of antigen in tissue odate consumption by collagen and by insulin. J Histochem cells; improvements in a method for the detection of antigen Cytochem 2:291-299. by means of fluorescent antibody. J Exp Med 1:1-13. Barrnett RJ (1953). The histochemical distribution of protein- DeWitt LM (1906). Morphology and physiology of the islets of bound sulfhydryl groups. J Natl Cancer Inst 13:905-925. Langerhans in some vertebrates. J Exp Med 8:193-238. Barrnett RJ, Marshall RB, Seligman AM (1955). Histochemical Diamare V (1899). Studii comparative sulle isoli di Langerhans de demonstration of insulin in the islets of Langerhans. pancreas. Internat Monatschr Anat Physiol 16:155-205. Endocrinology 57:419-438. Ehrhart M, Grube D, Bader MF, Aunis D, Gratzl M (1986). Barrnett RJ, Seligman AM (1952a). Demonstration of protein- Chromogranin A in the pancreatic islet: cellular and subcel- bound sulfhydryl and disulfide groups by two new histochem- lular distribution. J Histochem Cytochem 34:1673-1682. ical methods. J Natl Cancer Inst 13:215-216. Epple A (1967). A staining sequence for A, B and D cells of pan- Barrnett RJ, Seligman AM (1952b). Histochemical demonstration creatic islets. Biotech Histochem 42:53-61. of protein-bound sulfhydryl groups. Science 116:323-327. Erlandsen SL (1980). Types of pancreatic islet cells and their Barrnett RJ, Seligman AM (1954). Histochemical demonstration immunocytochemical identification. Monogr Pathol 21: of sulfhydryl and disulfide groups of protein. J Natl Cancer 140-155. Inst 14:769-803. Erlandsen SL, Hegre OD, Parsons JA, McEvoy RC, Elde RP Baskin DG (2014). ‘Fixation and tissue processing in immu- (1976). Pancreatic islet cell hormones distribution of cell nohistochemistry’ in McManus LM, Mitchell RN, (eds.) types in the islet and evidence for the presence of somatosta- Pathobiology of Human Disease. San Diego: Elsevier, pp. tin and gastrin within the D cell. J Histochem Cytochem 24: 3797-3806. 883-897. Baskin DG, Gorray KC, Fujimoto WY (1984). Evgenov NV, Medarova Z, Pratt J, Pantazopoulos P, Leyting Immunocytochemical identification of cells containing S, Bonner-Weir S, Moore A (2006a). In vivo imaging of Histochemical Identification of Islet Cell Types 557

immune rejection in transplanted pancreatic islets. Diabetes Hewitt SM, Baskin DG, Frevert CW, Stahl WL, Rosa-Molinar 55:2419-2428. E (2014). Controls for immunohistochemistry: the Evgenov NV, Medarova Z, Dai G, Bonner-Weir S, Moore A Histochemical Society’s standards of practice for validation (2006b). In vivo imaging of islet transplantation. Nat Med of immunohistochemical assays. J Histochem Cytochem 12:144-148. 62:693-697. Falck B, Hellman B (1963). Evidence for the presence of biogenic Jindal RM, Gray DW, Morris PJ (1993). The use of TSQ as an amines in pancreatic islets. Cell Mol Life Sci 19:139-140. islet-specific stain for purification of islets by fluorescence- Falkmer S, Knutson F, Voigt GE (1964). On the cobalt-concen- activated sorting. Transplantation 56:1282-1284. trating ability of pancreatic islet tissue. A histological, his- Kito H, Hosoda S (1977). Triple staining for simultaneous visu- tochemical, and autoradiographical investigation. Diabetes alization of cell types in islet of Langerhans of pancreas. 13:400-407. Successive application of argyrophil, aldehyde-fuchsin and Fujita T (1968). D cell, the third endocrine element of the pancre- lead-hematoxylin stains in a single tissue section. J Histochem atic islet. Arch Histol Jap 29:1-40. Cytochem 25:1019-1020. Garry DJ, Sorenson RL, Elde RP, Maley BE, Madsen A (1986). Kvistberg D, Lester G, Lazarow A (1966). Staining of insulin with Immunohistochemical colocalization of GABA and insulin in aldehyde fuchsin. J Histochem Cytochem 14:609-611. beta-cells of rat islet. Diabetes 35:1090-1095. Lacy PE (1957). Electron microscopic identification of different Garry DJ, Appel NM, Garry MG, Sorenson RL (1988). Cellular and cell types in the islets of Langerhans of the guinea pig, rat, subcellular immunolocalization of L-glutamate decarboxylase rabbit and dog. Anat Rec 128:255-267. in rat pancreatic islets. J Histochem Cytochem 36:573-580. Lacy PE, Davies J (1957). Preliminary studies on the demonstra- Goldsmith PC, Rose JC, Arimura A, Ganong WF (1975). tion of insulin in the islets by the fluorescent antibody technic. Ultrastructural localization of somatostatin in pancreatic islets Diabetes 6:354-357. of the rat. Endocrinology 97:1061-1064. Lacy PE, Davies J (1959). Demonstration of insulin in mammalian Gomori G (1939). A differential stain for cell types in the pancre- pancreas by the fluorescent antibody method. Stain Technol atic islets. Am J Pathol 15:497-499. 34:85-89. Gomori G (1941). Observations with differential stains on human Laguesse E (1893). Sur la formation des ilots de Langerhans dans islets of langerhans. Am J Pathol 17:395-406. le pancreas. Comptes Rend Soc Biol 5:819-820. Gomori G (1950). Aldehyde-fuchsin: a new stain for elastic tissue. Laguesse E (1895). Recherches sur l’histogenie du panréas chez le Am J Clin Pathol 20:665-666. mouton. J Anat Physiol 31:475-500. Gray DW, Millard PR, McShane P, Morris PJ (1983). The use of Lane MA (1908). The cytological characters of the areas of the dye neutral red as a specific, non-toxic, intra-vital stain of Langerhans. Am J Anat 7:409-422. islets of Langerhans. Brit J Exp Pathol 64:553-558. Liew CG, Shah NN, Briston SJ, Shepherd RM, Khoo CP, Dunne Greenwell MV, Nettleton GS, Feldhoff RC (1983). An investi- MJ, Moore HD, Cosgrove KE, Andrews PW (2008). PAX4 gation of aldehyde fuchsin staining of unoxidized insulin. enhances beta-cell differentiation of human embryonic stem Histochemistry 77:473-483. cells. PLoS One 3:e1783. Greider MH, Gersell DJ, Gingerich RL (1978). Ultrastructural Luft R, Efendic S, Hokfelt T, Johansson O, Arimura A (1974). localization of pancreatic polypeptide in the F cell of the dog Immunohistochemical evidence for the localization of soma- pancreas. J Histochem Cytochem 26:1103-1108. tostatin--like immunoreactivity in a cell population of the Grimelius L (1968a). A silver nitrate stain for alpha-2 cells in pancreatic islets. Med Biol 52:428-430. human pancreatic islets. Acta Soc Med Ups 73:243-270. Lukinius A, Wilander E, Westermark GT, Engstrom U, Grimelius L (1968b). The argyrophil reaction in islet cells of adult Westermark P (1989). Co-localization of islet amyloid poly- human pancreas: studies with a new silver nitrate procedure. peptide and insulin in the B cell secretory granules of the Acta Soc Med Upsal 73:271-294. human pancreatic islets. Diabetologia 32:240-244. Hefei W, Yu R, Haiqing W, Xiao W, Jingyuan W, Dongjun L Lukowiak B, Vandewalle B, Riachy R, Kerr-Conte J, Gmyr (2015). Morphological characteristics and identification of islet- V, Belaich S, Lefebvre J, Pattou F (2001). Identification like cells derived from rat adipose-derived stem cells cocultured and purification of functional human beta-cells by a new with pancreas adult stem cells. Cell Biol Int 39:253-263. specific zinc-fluorescent probe. J Histochem Cytochem Hegre OD, Leonard RJ, Erlandsen SL, McEvoy RC, Parsons JA, 49:519-528. Elde RP, Lazarow A (1976). Transplantation of islet tissue in Mason TE, Phifer RF, Spicer SS, Swallow RA, Dreskin RB the rat. Acta Endocrinol Suppl (Copenh) 205:257-281. (1969). An immunoglobulin-enzyme bridge method for local- Hellerstrom C, Hellman B (1960). Some aspects of silver impreg- izing tissue antigens. J Histochem Cytochem 17:563-569. nation of the islets of Langerhans in the rat. Acta Endocrinol Malaisse-Lagae F, Carpentier JL, Patel YC, Malaisse WJ, Orci L 35:518-532. (1977). Pancreatic polypeptide: a possible role in the regula- Hellman B, Hellerstrom C (1961). The specificity of the argyrophil tion of food intake in the mouse. Experientia 33:915-917. reaction in the islets of Langerhans in man. Acta Endocrinol McGadey J (1979). A staining sequence for the differentiation of 36:22-30. A-, B-, and D-cells of the islets of guinea-pig pancreas. Acta Hellman B, Wallgren A, Hellerstrom C (1962). Two types of islet Diabetolog Lat 16:243-246. alpha cells in different parts of the pancreas of the dog. Nature McNary WF (1954). Zinc-dithizone reaction of pancreatic islets. J 194:1201-1202. Histochem Cytochem 2:185-194. 558 Baskin

Minkowski O (1893). Untersuchungen über den Diabetes mellitus Sanger F, Tuppy H (1951b). The amino-acid sequence in the phe- nach Exstirpation des Pankreas. Arch Exp Pathol Pharmakol nylalanyl chain of insulin. I. The identification of lower pep- 31:85-189. tides from partial hydrolysates. Biochem J 49:463-481. Misugi K, Howell SL, Greider MH, Lacy PE, Sorenson GD Schäfer EA (1895). On internal secretion. BMJ 2:341-348. (1970). The pancreatic . Demonstration with peroxi- Schiebler TH, Schiessler S (1959). Über den Nachweis von Insulin dase-labeled antibody technique. Arch Pathol 89:97-102. mit den metachromatisch reagierenden Pseudoisocyaninen. Mosca L (1957). An experimental study of the cytology of pancre- Histochemie 1:445-465. atic islets. Q J Exp Physiol Cogn Med Sci 42:49-55. Schultze W (1900). Die Bedeutung der Langerhanschen Inseln im Mowry RW (1983). Selective staining of pancreatic beta-cell Pankreas. Arch Mikr Anat Entwickl 56:491-504. granules. Evolution and present status. Arch Pathol Lab Med Schweisthal MR, Clark LD, Shevell JL (1981). The staining of 107:464-468. pancreatic delta cells for light microscopy. Stain Technol Mowry RW, Kent SP (1988). Aldehyde pararosanilin (true alde- 56:25-28. hyde fuchsin) stains purified insulin, oxidized or not, follow- Scopsi L, Wang BL, Larsson LI (1986). Nonspecific immunocyto- ing adequate fixation with either formalin or Bouin’s fluid. chemical reactions with certain neurohormonal peptides and Stain Technol 63:311-323. basic peptide sequences. J Histochem Cytochem 34:1469- Mowry RW, Longley JB, Emmel VM (1980). Only aldehyde 1475. fuchsin made from pararosanilin stains pancreatic B cell Scott HR (1952). Rapid staining of beta cell granules in pancreatic granules and elastic fibers in unoxidized microsections: prob- islets. Stain Technol 27:267-268. lems caused by mislabelling of certain basic fuchsins. Stain Scott HR, Clayton BP (1953). A comparison of the staining affini- Technol 55:91-103. ties of aldehyde fuchsin and the Schiff reagent. J Histochem Meyer J, Bencosme SA (1965). The fine structure of normal rabbit Cytochem 1:336-352. pancreatic islet cells. Rev Can Biol 24:179-205. Shiroi A, Yoshikawa M, Yokota H, Fukui H, Ishizaka S, Tatsumi Ohta Y, Kosaka Y, Kishimoto N, Wang J, Smith SB, Honig G, Kim K, Takahashi Y (2002). Identification of insulin-producing H, Gasa RM, Neubauer N, Liou A, Tecott LH, Deneris ES, cells derived from embryonic stem cells by zinc-chelating German MS (2011). Convergence of the insulin and serotonin dithizone. Stem Cells 20:284-292. programs in the pancreatic β-cell. Diabetes 60:3208-3216. Sternberger LA, Hardy PH Jr, Cuculis JJ, Meyer HG (1970). The Okamoto K (1942). Biologische Untersuchungen der Metalle. VI. unlabeled antibody enzyme method of immunohistochemis- Histochemischer Nachweis einiger Metalle in den Geweben, try: preparation and properties of soluble antigen-antibody besonders in den Nieren, and deren Veranderungen. Tr Soc complex (horseradish peroxidase-antihorseradish peroxi- Pathol Jap 32:99-101. dase and its use in identification of spirochetes. J Histochem Opie EL (1901a). On the relation of chronic interstitial pancreatitis Cytochem 18:315-333. to the islands of Langerhans and to diabetes mellitus. J Exp Thomas TB (1937). Cellular components of the mammalian islets Med 5:397-428. f Langerhans. Am J Anat 63:31-57. Opie EL (1901b). The relation of diabetes mellitus to lesions of the Ustione A, Piston DW (2012). Dopamine synthesis and D3 recep- pancreas. Hyaline degeneration of the islands of Langerhans. tor activation in pancreatic β-cells regulates insulin secre- J Exp Med 5: 527-540. tion and intracellular [Ca(2+)] oscillations. Mol Endocrinol Orci L (1974). A portrait of the pancreatic B cell. Diabetologia 26:1928-1940. 10:163-187. von Mering J, Minkowski O (1890). Diabetes mellitus nach Orci L, Baetens D, Dubois MP, Rufener C (1975). Evidence for Pankreasextirpation. Archiv Exp Pathol Pharmakol 26: the D-cell of the pancreas secreting somatostatin. Hormone 371-378. Met Res 7:400-402. Wagner GF, McKeown BA, Popham DJ (1981). The autora- Pearse AG (1969). The cytochemistry and ultrastructure of poly- diographic localization of zinc within the pancreatic islets peptide hormone-producing cells of the APUD series and the of the rainbow trout, Salmo gairdneri. Histochemistry 72: embryologic, physiologic and pathologic implications of the 113-121. concept. J. Histochem. Cytochem 17: 303-313. Westermark GT, Westermark P (2013). Islet amyloid polypeptide Pearse AG, Polak JM (1971). Neural crest origin of the endocrine and diabetes. Curr Protein Pept Sci 14:330-337. polypeptide (APUD) cells of the gastrointestinal tract and Wierup N, Yang S, McEvilly RJ, Mulder H, Sundler F (2004). pancreas. Gut 12:783-788. Ghrelin is expressed in a novel endocrine cell type in devel- Pelletier G (1977). Identification of four cell types in the human oping rat islets and inhibits insulin secretion from INS-1 endocrine pancreas by immunoelectron microscopy. Diabetes. (832/13) cells. J Histochem Cytochem 52:301-310. 26:749-756. Wilander E, Westermark P (1976). On argyrophil reactions of Pictet RL, Rall LB, Phelps P, Rutter WJ (1976). The neural crest endocrine cells in the human fetal pancreas: a light and elec- and the origin of the insulin-producing and other gastrointes- tron microscopic study. Cell Tissue Res 68:33-43. tinal hormone-producing cells. Science 191:191-192. Williamson JR, Lacy PE (1959). Electron microscopy of islet cells Sakula A (1988). Paul Langerhans (1847-1888): a centenary trib- in alloxan-treated rabbits. AMA Arch Pathol 67:102-109. ute. J Royal Soc Med 81:414-415. Zacharovová K, Berková Z, Jirák D, Herynek V, Vancová M, Sanger F, Tuppy H (1951a). The amino-acid sequence in the phe- Dovolilová E, Saudek F (2012). Processing of superparamag- nylalanyl chain of insulin. 2. The investigation of peptides netic iron contrast agent ferucarbotran in transplanted pancre- from enzymic hydrolysates. Biochem J 49:481-490. atic islets. Contrast Media Mol Imaging 7:485-493.